Hrvoje> "Stephen J. Turnbull" <turnbull(a)sk.tsukuba.ac.jp&gt; writes: Hrvoje> Why do you keep repeating that? It breaks that kind of Hrvoje> code only for Asian use. First of all, it breaks it for all multilingual use. That means Croats who want to write in Latin-1 and Latin-2 at the same time, and pretty much all Israelis, Greeks, Arabs, and Russians, I bet. Second, I'm repeating it because as far as I know everybody who writes Emacs Lisp would prefer that their code function correctly in as many environments as possible, as long as it costs them nothing to write it that way. You feel that way, too, you've said so. Do you habitually write code _knowing_ that if the wrong somebody borrows it, their application will break? I thought not! There may be a few die-hards who would insist that the cost of fixing this problem be exactly zero, but I doubt that even you or Kyle are among them. The question is how high can it go? I would rather not break anybody's code at all. But code that uses int-char, explicitly or implicitly, is _already_ broken _by design_ unless either the code traps out-of-bounds data (fat chance, it's an optimization in the first place), or we do. Leaving it alone means breaking somebody's XEmacs someday. Why does it matter if that person is an Asian or a polyglot? Kyle's right: we owe our existing users and developers due consideration, and given the current distribution, that may mean compromising reliability for some existing users and many future ones. But Hope is out of Pandora's box; it is no longer possible to avoid some kind of compromise by ignoring the needs of Mule users. "They" is (some of) us, now; we can't go back. -- University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091 __________________________________________________________________________ __________________________________________________________________________ What are those two straight lines for? "Free software rules."